eNOS-knockout mice. Low dose: 30 mice received 1 mg/kg, 0.2 mg/ kg, or vehicle for 14 days before SAH, and motor function was assessed for 3 days after SAH. Post-SAH dose: 15 New Zealand white rabbits underwent autologous blood injection into the cisterna magna or sham surgery. Rabbits received simvastatin 60 mg/kg or vehicle for 3 days after SAH. Basilar artery lumen patency was assessed 72 hours after SAH. Human dose: 17 patients were randomized to simvastatin 80 mg/day or placebo for 14 days after SAH. Serum aspartate aminotransferase/alanine aminotransferase, creatine phosphokinase, von Willebrand factor (vWF), S-100a, tumor necrosis factor (TNF)-, and transcranial Doppler (TCD) velocities were recorded for 14 days. RESULTS: Simvastatin 20 mg/kg versus vehicle attenuated MCA vasospasm (percent lumen reduction, 27  18% versus 51  17%, P = 0.02), increased cerebrovascular eNOS protein threefold (P = 0.04), and reduced neurodeficits (P = 0.01) in wild-type but not in eNOS- knockout mice (percent lumen reduction, 46  20% versus 50  19%, P = 0.82). Simvastatin 1 mg/kg, not 0.2 mg/kg, reduced neurological deficits after SAH. In rabbits, post-SAH simvastatin 60 mg/kg atten- uated basilar artery vasospasm (percent lumen reduction, 16  5% versus 35  10%; P = 0.01). In patients, simvastatin 80 mg/d reduced serum vWF, S-100a, TNF, and TCD vasospasm (mean, 200 cm/s; 0% versus 38%, P = 0.08). No patients developed hepatitis/myositis. CONCLUSION: High-dose simvastatin before SAH attenuated va- sospasm and neurological deficits in mice. This effect persisted when simvastatin was given after SAH or at lower doses approved for human use but was lost in eNOS-knockout mice, suggesting an eNOS- dependent mechanism of action. Simvastatin was safe in patients with SAH and may be an effective agent in the management of SAH. 765 Genome-wide Linkage Analysis of Intracranial Aneurysms Brian V. Nahed, B.S., Abigail A. Hawkins, B.A., Michael L. Diluna, M.D., Askin Seker, M.D., Bulent Guclu, M.D., Andrea Chamberlain, B.S.N., C.C.R.N., R.N., Mathew W. State, M.D., Ph.D., Murat Gunel, M.D. INTRODUCTION: Genetic risk factors, in addition to environmen- tal risk factors, contribute to the formation and/or rupture of cerebral aneurysms. Previous genome-wide linkage studies suggest linkage at 5q22-q31, 7q11, 14q22, and 19q13.3 with candidate genes such as elastin, COL3A1 (ch 2q31), fibrillin (ch 15q21), and endoglin (intron 7). However, attempts to reproduce these results failed, indicating the limitations inherent in nonparametric and candidate gene analysis. In this study, genome-wide linkage analysis and parametric methods were used to identify the gene(s) leading to intracranial aneurysm (ICA). METHODS: Since 1994, we have prospectively screened over 3000 patients with ICAs and their relatives, identifying 142 multiplex fam- ilies with 345 affected patients. Twenty-one of these families have more than four affected members, and four have more than seven affected members. Using Affymetrix genechips containing over 10,500 single-nucleotide polymorphisms, we performed genome-wide link- age analysis on our largest kindreds to identify regions of linkage. Once regions were identified, we fine-mapped these loci by using microsatellite markers to narrow these linked intervals. Subsequent candidate gene analysis provided novel genes of interest. RESULTS: Genome-wide linkage analysis using Genechips iden- tifed a total of eight loci linked to ICA phenotype that gave near maximum theoretical logarithm of the odds scores. Further analysis using microsatelite markers excluded five of these regions, leaving three loci possibly contributing to the formation of ICA. We are now performing mutational analysis within these three regions. CONCLUSION: Parametric methods successfully identified novel chromosomal regions linked to ICA. Subsequent fine-mapping of these regions has confirmed linkage at our novel sites and excluded other regions, particularly those previously thought to be ICA-related. This might be caused by genetic heterogeneity of ICA or because of inherent limitations of nonparametric studies. Finally, gene screening and candidate gene analysis within these linked intervals have re- vealed at least two novel candidate genes involved in angiogenesis, cell adhesion, and remodeling. 766 Vasospasm Probability Index: A Combination of Transcranial Doppler Velocities, Cerebral Blood Flow, and Clinical Risk Factors to Predict Cerebral Vasospasm after Subarachnoid Aneurysm Hemorrhage Nestor Gonzalez, M.D., Thomas Glenn, Ph.D., John Boscardin, Ph.D., Fernando Vin ˜ uela, M.D., Neil Martin, M.D. INTRODUCTION: The purpose of this study was to evaluate trans- cranial Doppler (TCD) velocities, Lindegaard ratio (LR), and spasm index (SI) in the diagnosis of cerebral vasospasm and to create a vasospasm probability index (VPI) to improve the vasospasm diag- nosis accuracy after subarachnoid hemorrhage (SAH). METHODS: A total of 795 patients with SAH underwent TCDs between April 1998 and January 2000. Of these, 154 patients had angiographies and 75 had 133 Xe cerebral blood flow studies the same day. Seven cases were excluded because of poor sonographic window. Factors analyzed included age, sex, Hunt and Hess and Fisher grades, day after SAH, treatment day, treatment type (clipping, embolization, or conservative), and tobacco and hypertension history. Contingency tables were constructed to calculate tests accuracy. Logistic regression was used to evaluate the predictive factors, and their coefficients were integrated in the VPI. RESULTS: Forty-one women (60.3%) and 27 men, age 35 to 84 years (mean, 58 yr) were included. Eighteen (26.5%) had symptomatic va- sospasm. Mean middle cerebral artery-TCD velocities in patients with vasospasm was 147.9 cm/s (standard deviation, 57.0), versus 83.1cm/s (standard deviation, 41.1) (P  0.0001). Average LR was 5.14 (standard deviation, 2.48) for arteries with vasospasm, versus 2.42 (standard deviation, 1.01) (P  0.0001). Mean SI for vasospasm cases was 5.11 (standard deviation, 2.11) versus 2.32 (standard deviation, 1.05) (P  0.0001). Thirty-three patients (48.5%) had angiographic vasospasm. Mean middle cerebral artery-TCD velocity was 142.7 cm/s (standard deviation, 62.0), versus 77.5 cm/s (standard deviation, 32.4) (P  0.0001). Mean LR in arteries with vasospasm was 4.76 (standard deviation, 2.35), versus 2.29 (standard deviation, 0.79) (P  0.0001). Mean SI in angiographic vasospasm was 4.49 (standard devi- ation, 2.17) versus 2.27 (standard deviation, 1.01) (P  0.0001). No significant differences were observed between surgical, endovascular, and conservative treatment or in patients with history of hypertension or smokers. Sensitivity, specificity, and predictive values are shown in Table 1. After logistic regression analysis, significant predictors of vasospasm were Fisher grade, Hunt and Hess grade, LR, and SI. A ABSTRACTS OF OPEN PAPERS 474 | VOLUME 55 | NUMBER 2 | AUGUST 2004 www.neurosurgery-online.com Downloaded from https://academic.oup.com/neurosurgery/article-abstract/55/2/474/2744431 by guest on 21 March 2018